PREENCHIMENTO DE FALHAS DE SÉRIES DE DADOS CLIMÁTICOS UTILIZANDO REDES P2P

• Ano de defesa: 2012
• Autor(a) principal: Schmitke, Luiz Rafael
• Orientador(a): Senger, Luciano José
• Banca de defesa: Vaz, Maria Salete Marcon Gomes , Virgens Filho, Jorim Sousa das
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: UNIVERSIDADE ESTADUAL DE PONTA GROSSA
• Programa de Pós-Graduação: Programa de Pós Graduação Computação Aplicada
• Departamento: Computação para Tecnologias em Agricultura
• País: BR
• Palavras-chave em Português: Redes P2P; correção; Dados Climáticos
• Palavras-chave em Inglês: P2P Networks; correct; Climate Data
• Área do conhecimento CNPq: CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: http://tede2.uepg.br/jspui/handle/prefix/150

Agriculture is an activity where the weather has more impact, influencing techniques and crops employed. Much of the agricultural productivity is affected by climatic conditions that are created by natural factors and are not likely to control. Although you can’t control the weather, we can predict it, or even simulate their conditions to try minimize its impact on agriculture. To be able to make these predictions and simulations are necessary data collected from weather stations that can be conventional or automatic and must be without gaps or abnormal data. Most of these errors are caused by signal interference, disconnection, oxidation of cables and spatio-temporal variation of climate which consequently end up generating those problems at the climates bases. Thus, this research work has as main objective to create a model capable of correcting gaps in climate databases, observing that not to correct abnormal observations or replace statistical methods for the same purpose. Therefore a model was created to correct the gaps in weather data between stations using the P2P architecture. With this model, an application was created to test its performance to correct the gaps. Also to perform the tests were used bases in the cities of Ponta Grossa, Fernandes Pinheiro and Telêmaco Borba provided by Instituto Tecnológico SIMEPAR, and bases of the cities of Castro, Carambeí, Pirai do Sul and Tibagi provided by Fundação ABC, which are collected daily on automatic stations. As a result it was observed that the performance of P2P correction model was satisfactory when compared to the simulator used in the tests, with lower results only in February, which corresponds to the period of summer, to the autumn, winter and spring the model P2P was better than simulated. Although it was found that the number of stations participating in the network at the time of correcting influences the results, and the higher it is, the better the results obtained with the correcting.